Ductal carcinoma in situ (DCIS) of the breast is a heterogenous disease with varying clinical outcomes. Current classification of DCIS is based on cytological and architectural features. Although some of these features correlates with the risk of local recurrence, they can be determined only with moderate consistency and none of them is able to accurately predict the development of invasive disease. Therefore the identification of novel molecular markers that correlate with the clinical behavior of DCIS is of utmost importance. The specific goals of this proposal are aimed at the identification and evaluation of such markers using the combination of comprehensive gene expression analysis and tissue microarrays. In order to define genes implicated in DCIS, we have analyzed the gene expression profiles of DCIS and normally mammary epithelium using SAGE (Serial Analysis of Gene Expression) and identified over 300 differentially expressed transcripts. To further analyze how reproducibly these genes are differentially expressed we propose to generate tissue microarrays from hundreds of samples of benign breast tissue and in situ and invasive breast carcinomas. These tissue microarrays will then be used to rapidly screen DCIS specific transcripts using in situ hybridization and immunohistochemistry. The clinical usefulness of genes with consistently different expression in DCIS will be evaluated using a tissue microarray generated from DCIS cases with known clinical outcome. Identification of clinically useful molecular markers that would predict the behavior of DCIS could lead to the individualized treatment of patients diagnosed with this disease.